connections between combinatorics and commutative...

Connections between Combinatorics and

Commutative Algebra

Ayesha Asloob Qureshi

Sabancı Universitesi

Andalas University, Padang

November 8-9, 2018

Combinatorial commutative algebra combines the abstract

methods of algebra, geometry and topology with the more

intuitive ones of combinatorics.

Combinatorial commutative algebra combines the abstract

methods of algebra, geometry and topology with the more

intuitive ones of combinatorics.

Several applications in Chemistry, Biology, Statistics, Computer

Sciences..

Following the work Richard Stanley, in the late 1970’s a new

and exciting trend started in commutative algebra, namely, the

combinatorial study of squarefree monomial ideal.

In 1966, H. Anand, V. C. Dumir, and H. Gupta investigated a

combinatorial distribution problem and formulated some

conjectures on the number of solutions.

In 1966, H. Anand, V. C. Dumir, and H. Gupta investigated a

combinatorial distribution problem and formulated some

conjectures on the number of solutions.

Suppose that n distinct objects, each available in r identical

copies, are distributed among n persons in such a way that

each person receives exactly r objects. What can be said about

the number H(n, r) of such distributions?

The problem can be restructed as:

Let aij denote the number of copies of object i that person j

receives, then A = (aij) ∈ Z n×n is an n × n matrix such that the

sum of each row and column of A is r .

n∑

k=1

aik =

n∑

l=1

alj = r , for i , j = 1, . . . ,n.

Then H(n, r) is the number of such matrices A.

4 9 2

3 5 7

8 1 6

15 15 15

15

15

15

15

Conjecture proposed by Anand, Dumir and Gupta

Conjecture proposed by Anand, Dumir and Gupta

◮ there exists a polynomial Pn(r) of degree (n − 1)2such that

H(n, r) = Pn(r) for all r > n; in particular Pn(r) = 0 for

r = 1, . . . ,n − 1

Let Mn be the set of solutions. Then Stanley observed the

following:

◮ Mn has an algebraic sturucture of a submnoid of Zn×n+ .

◮ The monoid algebra K [Mn] has Krull dimension (n−1)2 +1

◮ By using the theory of Hilbert functions, one see that

conjectures hold.

The d -dimensional cyclic polytope with n vertices is the convex

hull

C(n,d) := conv .hull{x(t1), x(t2), . . . , x(tn)}

of n > d ≥ 2 distinct points x(ti) on the moment curve.

The d -dimensional cyclic polytope with n vertices is the convex

hull

C(n,d) := conv .hull{x(t1), x(t2), . . . , x(tn)}

of n > d ≥ 2 distinct points x(ti) on the moment curve.

The moment curve in Rd is

x : R→ Rd , x(t) := (t , t2, . . . , td ).

Theorem (Upper bound ”Conjecture” )

The cyclic polytopes have the maximum possible number of

faces for a given dimension and number of vertices. If ∆ is a

simplicial sphere of dimension d − 1 with n vertices, then

fi(∆) ≤ fi(C(n,d)), for i = 0,1, . . . ,d − 1

The conjecture was proposed in 1957.

Proved in 1970: for simplicial polytopes by Peter McMullen.

Proved in 1975: for simplicial spheres by Richard P. Stanley.

Following the work Richard Stanley, in the late 1970’s a new

and exciting trend started in commutative algebra, namely, the

combinatorial study of squarefree monomial ideal.

◮ Let K be a field.


◮ S = K [x1, . . . , xn] be the polynomial ring in n variables.

Simply speaking, S is the set of all polynomials in variables

x1, . . . , xn with coefficients in K .





◮ A monomial in S is a product of variables,

xa1

1 xa2

2 . . . xann , (a1,a2, . . . ,an) ∈ N

for example, x2x24 x5.





◮ A monomial in S is a product of variables,

xa1

1 xa2

2 . . . xann , (a1,a2, . . . ,an) ∈ N

for example, x2x24 x5.

◮ A squarefree monomial is a monomial in which no variable

can appear twice, for example x2x4x5.

Let A = {u1, . . . ,ur} be a set of r squarefree monomials in S.

Typically, we can associate two algebraic structures of

combinatorial nature related to A.

◮ The sqaurefree monomial ideal generated by u1, . . . ,ur in

S.





S.

Recall that I ⊂ S is an ideal of S if for all x , y ∈ I and r ∈ S,

we have x + y ∈ I and rx ∈ I.





S.



◮ The toric ring, K [A] = K [u1, . . . ,um], which is monomial

subring of S.





S.




subring of S.

Let T = K [t1, . . . , tm]





S.




subring of S.

Let T = K [t1, . . . , tm] and define the surjective homomorphism

φ : T → K [A]

by setting φ(ti) = ui for i = 1, . . . ,m.





S.




subring of S.


φ : T → K [A]

by setting φ(ti) = ui for i = 1, . . . ,m. The toric ideal of A is the

kernel of φ.





S.




subring of S.


φ : T → K [A]

by setting φ(ti) = ui for i = 1, . . . ,m. The toric ideal of A is the

kernel of φ. In other words, the toric ideal of A is the defining

ideal of the toric ring K [A].

Let [n] = {1, . . . ,n}.

Let [n] = {1, . . . ,n}.

A simplicial complex ∆ on the vertex set [n] is a collection of

subsets in 2[n] with the following property:

F ∈ ∆ and G ⊂ F ⇒ G ∈ ∆.

Let [n] = {1, . . . ,n}.



F ∈ ∆ and G ⊂ F ⇒ G ∈ ∆.

The elements of ∆ are called faces of ∆.

Let [n] = {1, . . . ,n}.



F ∈ ∆ and G ⊂ F ⇒ G ∈ ∆.

The elements of ∆ are called faces of ∆.

dim∆ = max{|F | : F ∈ ∆} − 1

Example: Let ∆ =< {1,2,3}, {3,4,5}, {2,5}, {1,2}, {2,3},{1,3}, {3,4}, {3,5}.{4,5}, {1}, {2}, {3}, {4}, {5} >.

dim∆ = 2

Figure: Simplicial complex

Given a simplicial complex ∆ on vertex set [n], one can

associate a squarefree monomial ideal in a following way:

◮ Fix a field K . Let S = K [x1, x2, . . . , xn].




◮ Take all non-faces of ∆. A non-face of ∆ is subset

{i1, i2, . . . , it} ∈ 2[n] such that {i1, i2, . . . , it} /∈ ∆.






◮ To each non-facet {i1, i2, . . . , it} of ∆, attach a monomial

xi1xi2 . . . xit .







xi1xi2 . . . xit .

◮ I∆ = (xi1xi2 . . . xit : {i1, i2, . . . , it} /∈ ∆). Note that, because

of the algebraic structure of I∆, it is enough to consider to

the minimal non-faces.







xi1xi2 . . . xit .




◮ The ideal I∆ is called Stanley-Reisner ideal of ∆.







xi1xi2 . . . xit .




◮ The ideal I∆ is called Stanley-Reisner ideal of ∆.

◮ K [∆] = S/I∆ is called the Stanely-Reisner ring of ∆

Example: The minimal non-faces of ∆ are {1,4}, {1,5}, {2,4}and {2,3,5}. Then

I∆ = (x1x4, x1x5, x2x4, x2x3x5)

.

◮ Given a simplicial complex, one attach a simplicial complex

to it.

◮ Given a squarefree monomial ideal, one can attach a uniqe

simplicial complex to it.

◮ Stanley-Riesner correspondence

sqaurefree monomial ideals←→ simplicial complexes.

Every simplicial complex has a geometric realization.


A k-simplex σ ⊂ Rn is the convex hull of k + 1 affinely

independent points.



independent points.

Geometrically, simplicial complexes are collections of simplices

glued together by their faces.



independent points.

Geometrically, simplicial complexes are collections of simplices

glued together by their faces.

A geometric simplicial complex ∆ is a collection of simplices of

Rn such that whenever σ, σ′ ∈ ∆, one has

◮ If τ ⊂ σ, then τ ∈ ∆.

◮ σ ∩ σ is a face of both σ and σ′.

Recall that, dim∆ = max{|F | : F ∈ ∆} − 1

TheoremLet ∆ be a simplical complex on n vertices. Then

dim K [∆] = dim∆+ 1.

(Graph theory←→ Commutative algebra)

Let G be a simple finite graph with vertex set V (G) and edge

set E(G). Then the edge ideal associated with G is

I(G) = (xi xj : {i , j} ∈ E(G)) ⊂ K [x1, x2, . . . , xn]

Let G be the following graph on vertex set V (G) = {1,2,3,4}.

Then I(G) = (x1x2, x1x3, x1x4, x2x4, x3x4).

E(G) = {{1,2}, {1,3}, {1,4}, {2,4}, {3,4}}

A graph G is chordal if every cycle of four or more vertices of G

has a chord.

Theorem (Frobergs Theorem)

Let G be a graph. Then I(G) has a linear resolution if and only

if Gc is a chordal graph.

A graph G is chordal if every cycle of four or more vertices of G

has a chord.

Theorem (Frobergs Theorem)

Let G be a graph. Then I(G) has a linear resolution if and only

if Gc is a chordal graph.

Let I ⊂ S be a monomial ideal. Then the ideal I has a d-linear

resolution, if I has the following minimal graded free resolution:

0→ Sβt (−(d + t))→ · · · → Sβ1(−(d + 1))→ Sβ0(−d)→ I → 0

.

A colouring of a graph G is an assignment of a colour to each

vertex so that adjacent vertices, i.e., vertices joined by an edge

receive different colours. The chromatic number of a graph G,

denoted χ(G), is the minimum number of colours needed to

colour G.

Colouring is a core topic in graph theory. It has many practical

applications, including the problem of scheduling.

Colouring is a core topic in graph theory. It has many practical

applications, including the problem of scheduling.

Suppose we want to schedule a set of exams. Represent each

exam by a vertex, and join two vertices if there is a student who

must write both exams. For example, suppose we end up with

the graph:

We need 3 colors.

A subset W ⊂ V (G) is a vertex cover if W ∩ e 6= φ for all

e ∈ E(G). A vertex cover W is a minimal vertex cover if no

proper subset of W is a vertex cover.




xw =∏

xi∈W

xi




xw =∏

xi∈W

xi

J(G) = (xw : W is a minimal vertex cover of G)

Theorem (C.A. Francisco, H.T. Ha, and A. Van Tuyl)

χ(G) = min{d : (x1 . . . xn)d−1 ∈ J(G)d}.

Thank you

connections between combinatorics and commutative...

Documents